Artificial intelligence and its contribution to construction
09 September, 2019
What actually is artificial intelligence?
Artificial intelligence is based on the ability of a machine to analyse a situation, draw conclusions from it and act accordingly. AI technologies rely simultaneously on Big Data and dramatic increases in hardware computing power. They define algorithms based on identifying significant recurrences and integrating learning methods.
In parallel with the first public-facing applications, such as voice recognition and visual identification, every sector of industry is reconstructing and adapting AI according to meet its own specific needs and environment.
At VINCI Construction, we see artificial intelligence as primarily benefiting our people.
In the VINCI Construction context, the first application of artificial intelligence could focus on the risks inherent in construction by relieving the operators of all repetitive processes and automating them. “At VINCI Construction, we see AI as primarily benefiting our people. It doesn’t change the nature of what we do, but helps us to refocus on our core business”, explains Djamil Yahia-Ouahmed, Head of Data & AI at VINCI Construction.
No artificial intelligence without multiple data sources and accurate data collection
Data collection and organisation is a basic requirement for any AI application. But the fact remains that on construction sites, even with the more than promising advent of Building Information Modelling, data often remain scarce and/or fragmented. Djamil Yahia-Ouahmed: “The most important consideration is to think about what could be done with the data we have, rather than imagining an immediately applicable artificial intelligence solution”.
Nevertheless AI could help builders in the near future. Examples of possible applications include:
- analysing of highly structured documents, such as calls for tender. The opportunity to identify dates, entities and specific contractual clauses would facilitate the tender preparation process
- recognition of tasks and the automated entry of the relevant data into the appropriate programmes to provide automated reporting as work progresses on site would save time
- voice-activated virtual assistants would simplify the same reporting process, thereby ensuring more effective monitoring of production and unforeseen on-site events
- analysis of site-related contractual documents, which would then be transcribed using an interactive knowledge graph. This innovative method of data visualisation provides interactive links between contributors, entities and specific topics. The result is the ability to identify cause-and-effect relationships that would otherwise have been buried deep in the thousands of documents generated by construction projects.
Examples like these allow us to imagine other applications of AI for construction sites, especially in terms of predictive technologies. Algorithms can help to predict the future trend of the construction site operating account by analysing the time lag between schedules, the stage of progress and the impact on overall progress of incidents and complaints. By cross-referencing all available data, algorithms should then be able to calculate a percentage accident risk and notify site managers at the same time. Similarly, predictive maintenance could be driven by algorithms that generate machine maintenance schedules based not on intervals of weeks or months, but rather on actual usage data and on-site conditions. This would enable the transition away from something that is still programmed systematically and cyclically and towards dynamic resource allocation.
It is likely that we will soon see autonomous trucks in use on earthworks projects, even if they are initially confined to closed circuits with no human presence. The lack of space on urban construction sites would prohibit their use on these projects. Nevertheless, it’s important to know that AI is now capable of calculating a route to ensure that vehicles and pedestrians never meet!
But it’s primarily in safety that artificial intelligence is already showing promising advances and achievements. Video-based ‘guardian angel’ systems will be able to use image recognition to come to the assistance of someone working on scaffolding under potentially unsafe conditions. It can also detect whether someone is wearing personal protective equipment or not, and therefore - for example - prevent anyone not wearing a dosimeter from entering a nuclear industry worksite.
Maintaining control of your data as you step into the future
Tech giants, AI specialists and startups offering practical applications for the construction sector; this market is attracting more entrants all the time, each of which may play its own role in supporting VINCI Construction and its subsidiary companies to meet their artificial intelligence needs.
With support from Leonard, the laboratory set up to explore the future of the VINCI Group, and IT departments staffed increasingly by data scientists, AI architects and other newly emerging specialists, AI is assuming its rightful place as another valuable construction industry tool.